8 2. Backgroundan accurate signal strength map(SSM). A SSM represents the distribution of RSSIvalues from an AP over the space of an indoor scene. Therefore, the two approachesare distinguished by which means these RSSI values are obtained. The first approach,named empirical radio propagation model, is based on the technique to collect asignificant amount of real world RSSI measurements, so that the propagation modelcan afterwards predict RSSI estimates for arbitrary locations. It is crucial for thisapproach to gather enough information about the interesting zones of the scene.Furthermore, the data should preferably be collected homogeneously, for example,by applying a grid to the location space. It can be seen, that depending on theresolution of this grid, the construction of an empirical propagation model can be alaborious undertaking. Additionally, this approach becomes even more expensive ifchanges in the environment happen, for example by relocating APs or reorganisationof furniture. Such changes make a recalibration of the propagation model mandatory.Due to the expensive nature of the empirical approach, the research in this areas hasbeen focused on the alternate idea to create the sought RSSI distributions artificiallyby reasoning about the rules of radio propagation. Therefore, the class of thesemodels is named analytical radio propagation models. The most basic one is givenby assuming an idealized free-space environment and the corresponding quadraticalsignal power loss with respect to the distance between the current location andthe sender. This radio propagation model is called ideal path-loss model and hastherefore been ranked lowest in figure 2.1. The first adaptation to the environmentconditions is done in the general path-loss model by assuming a linearly elevatedquadratical power loss. The linear coefficient has to be determined empirically andcan be assumed to be higher for scenes with more Non Line Of Sight(NLOS) thanLOS conditions as more obstacles in the scene lead to a higher probability of signalabsorptions.Therefore, the analytical models needs to be adjusted with empirical estimated pa-rameters as well. These unknown parameters of the analytical propagation modelsare called free parameters. Another parameter driven analytical model has beendescribed in the RADAR system[2]. The presented model is the so called WallAttenuation Factor(WAF) model. The basic assumption is given by an assumedconstant signal decay at each obstacle intersection on the straight line between thesender and the simulated location in space. The model is easily enhanced to simulatedifferent types of obstacles, and will therefore be called multi-material WAF. Thefree material parameters of the multi-material WAF have also to be found empiri-cally. And the only physical effect, that the WAF simulates is an absorption of thesignalAn alternative approach, the dominant path model[1] adds the simulation of signalreflection. The change of direction of the signal at a material intersection is computedand the signals on the dominant paths(i.e. the signals with the maximum power)are traced until exhaustion. Their aggregated information of the traced paths willbe used as the basis for the SSM.As a general rule, more sophisticated radio propagation models can be obtainedby simulating more of the physical effects that influence the propagation of radiosignals. Such effects are especially found at the transition boundaries of opticalmedia, for example between air and solid material. The following physical effectscan be considered: